1 from __future__ import unicode_literals
15 ('>>', operator.rshift),
16 ('<<', operator.lshift),
20 ('/', operator.truediv),
23 _ASSIGN_OPERATORS = [(op + '=', opfunc) for op, opfunc in _OPERATORS]
24 _ASSIGN_OPERATORS.append(('=', lambda cur, right: right))
26 _NAME_RE = r'[a-zA-Z_$][a-zA-Z_$0-9]*'
29 class JSInterpreter(object):
30 def __init__(self, code, objects=None):
35 self._objects = objects
37 def interpret_statement(self, stmt, local_vars, allow_recursion=100):
38 if allow_recursion < 0:
39 raise ExtractorError('Recursion limit reached')
43 stmt_m = re.match(r'var\s', stmt)
45 expr = stmt[len(stmt_m.group(0)):]
47 return_m = re.match(r'return(?:\s+|$)', stmt)
49 expr = stmt[len(return_m.group(0)):]
52 # Try interpreting it as an expression
55 v = self.interpret_expression(expr, local_vars, allow_recursion)
56 return v, should_abort
58 def interpret_expression(self, expr, local_vars, allow_recursion):
61 if expr == '': # Empty expression
64 if expr.startswith('('):
66 for m in re.finditer(r'[()]', expr):
72 sub_expr = expr[1:m.start()]
73 sub_result = self.interpret_expression(
74 sub_expr, local_vars, allow_recursion)
75 remaining_expr = expr[m.end():].strip()
76 if not remaining_expr:
79 expr = json.dumps(sub_result) + remaining_expr
82 raise ExtractorError('Premature end of parens in %r' % expr)
84 for op, opfunc in _ASSIGN_OPERATORS:
86 (?P<out>%s)(?:\[(?P<index>[^\]]+?)\])?
88 (?P<expr>.*)$''' % (_NAME_RE, re.escape(op)), expr)
91 right_val = self.interpret_expression(
92 m.group('expr'), local_vars, allow_recursion - 1)
94 if m.groupdict().get('index'):
95 lvar = local_vars[m.group('out')]
96 idx = self.interpret_expression(
97 m.group('index'), local_vars, allow_recursion)
98 assert isinstance(idx, int)
100 val = opfunc(cur, right_val)
104 cur = local_vars.get(m.group('out'))
105 val = opfunc(cur, right_val)
106 local_vars[m.group('out')] = val
113 r'(?!if|return|true|false)(?P<name>%s)$' % _NAME_RE,
116 return local_vars[var_m.group('name')]
119 return json.loads(expr)
124 r'(?P<var>%s)\.(?P<member>[^(]+)(?:\(+(?P<args>[^()]*)\))?$' % _NAME_RE,
127 variable = m.group('var')
128 member = m.group('member')
129 arg_str = m.group('args')
131 if variable in local_vars:
132 obj = local_vars[variable]
134 obj = self._objects.setdefault(
135 variable, self.extract_object(variable))
139 if member == 'length':
143 assert expr.endswith(')')
149 self.interpret_expression(v, local_vars, allow_recursion)
150 for v in arg_str.split(',')])
152 if member == 'split':
153 assert argvals == ('',)
156 assert len(argvals) == 1
157 return argvals[0].join(obj)
158 if member == 'reverse':
159 assert len(argvals) == 0
162 if member == 'slice':
163 assert len(argvals) == 1
164 return obj[argvals[0]:]
165 if member == 'splice':
166 assert isinstance(obj, list)
167 index, howMany = argvals
169 for i in range(index, min(index + howMany, len(obj))):
170 res.append(obj.pop(index))
173 return obj[member](argvals)
176 r'(?P<in>%s)\[(?P<idx>.+)\]$' % _NAME_RE, expr)
178 val = local_vars[m.group('in')]
179 idx = self.interpret_expression(
180 m.group('idx'), local_vars, allow_recursion - 1)
183 for op, opfunc in _OPERATORS:
184 m = re.match(r'(?P<x>.+?)%s(?P<y>.+)' % re.escape(op), expr)
187 x, abort = self.interpret_statement(
188 m.group('x'), local_vars, allow_recursion - 1)
190 raise ExtractorError(
191 'Premature left-side return of %s in %r' % (op, expr))
192 y, abort = self.interpret_statement(
193 m.group('y'), local_vars, allow_recursion - 1)
195 raise ExtractorError(
196 'Premature right-side return of %s in %r' % (op, expr))
200 r'^(?P<func>%s)\((?P<args>[a-zA-Z0-9_$,]+)\)$' % _NAME_RE, expr)
202 fname = m.group('func')
204 int(v) if v.isdigit() else local_vars[v]
205 for v in m.group('args').split(',')])
206 self._functions.setdefault(fname, self.extract_function(fname))
207 return self._functions[fname](argvals)
209 raise ExtractorError('Unsupported JS expression %r' % expr)
211 def extract_object(self, objname):
214 (r'(?:var\s+)?%s\s*=\s*\{' % re.escape(objname)) +
215 r'\s*(?P<fields>([a-zA-Z$0-9]+\s*:\s*function\(.*?\)\s*\{.*?\}(?:,\s*)?)*)' +
218 fields = obj_m.group('fields')
219 # Currently, it only supports function definitions
220 fields_m = re.finditer(
221 r'(?P<key>[a-zA-Z$0-9]+)\s*:\s*function'
222 r'\((?P<args>[a-z,]+)\){(?P<code>[^}]+)}',
225 argnames = f.group('args').split(',')
226 obj[f.group('key')] = self.build_function(argnames, f.group('code'))
230 def extract_function(self, funcname):
233 (?:function\s+%s|[{;,]%s\s*=\s*function|var\s+%s\s*=\s*function)\s*
234 \((?P<args>[^)]*)\)\s*
235 \{(?P<code>[^}]+)\}''' % (
236 re.escape(funcname), re.escape(funcname), re.escape(funcname)),
239 raise ExtractorError('Could not find JS function %r' % funcname)
240 argnames = func_m.group('args').split(',')
242 return self.build_function(argnames, func_m.group('code'))
244 def call_function(self, funcname, *args):
245 f = self.extract_function(funcname)
248 def build_function(self, argnames, code):
250 local_vars = dict(zip(argnames, args))
251 for stmt in code.split(';'):
252 res, abort = self.interpret_statement(stmt, local_vars)